Estimating household income to monitor and evaluate public investment programs in Sub-Saharan Africa

However, accurate assessment of rural household income is time consuming and costly. Using an expenditure-based income measure, data on actual household expenditures per capita obtained from various national surveys for 28 Sub-Saharan African countries, this study used proxy indicators to estimate regression models and then predict and analyze changes in household income per capita between 1985 and 2006. Over the 20-year period, the study predicted annual average real household monthly income per capita at $78 in 1993 international dollars. South Africa was ahead of the group of countries at $225, followed by C�te d"Ivoire and Lesotho at $117 and $91, respectively. Predictions for Nigeria and Zambia were the worst at $28 and $39, respectively. Looking at changes in income over time, Burkina Faso, C�te d"Ivoire, Uganda, Senegal, Mauritania, and Ghana (in declining order) experienced consistent positive growth. In contrast, Zambia, Kenya, and Lesotho showed declining trends, averaging �2.7 percent, �2.0 percent, and �1.3 percent per year, respectively, over the 20-year period. The latter results were not surprising given the low and sometimes negative growth rates in real GDP per capita and real agricultural value added per worker over the same period for those countries. The predicted trends were also consistent with observed trends in poverty and hunger, suggesting that the methodology is a useful and least-cost approach for monitoring household incomes to support evaluation of public investment programs. (Author abstract)